1 |
On Tue, 20 Jan 2015 10:40:17 +0100 |
2 |
Michał Górny <mgorny@g.o> wrote: |
3 |
|
4 |
> # Copyright 1999-2014 Gentoo Foundation. |
5 |
> # Distributed under the terms of the GNU General Public License v2 |
6 |
> # $Header$ |
7 |
> |
8 |
> # Whenever the flag name does not correspond to /proc/cpuinfo flags, |
9 |
> # please put the cpuinfo flag in square brackets. If more than one |
10 |
> flag # is specified, any of them will enable the respective USE flag. |
11 |
> |
12 |
|
13 |
nice |
14 |
|
15 |
|
16 |
now, letmegrepitforyou: |
17 |
|
18 |
|
19 |
> avx - Adds support for Advanced Vector Extensions instructions |
20 |
|
21 |
media-video/ffmpeg:avx - Adds support for Advanced Vector Extensions |
22 |
instructions (Intel Sandy Bridge, AMD Bulldozer and later chips) |
23 |
|
24 |
> avx2 - Adds support for Advanced Vector Extensions 2 instructions |
25 |
|
26 |
media-video/ffmpeg:avx2 - Adds support for Advanced Vector Extensions 2 |
27 |
instructions (Intel Haswell and later chips) |
28 |
|
29 |
> fma3 - Use the Fused Multiply Add 3 instruction set ([fma] in cpuinfo) |
30 |
|
31 |
media-video/ffmpeg:fma3 - Enables FMA3 optimizations: AMD processors |
32 |
starting with Piledriver architecture and Intel Haswell based |
33 |
processors or later. |
34 |
|
35 |
> fma4 - Use the Fused Multiply Add 4 instruction set |
36 |
|
37 |
media-video/ffmpeg:fma4 - Enables FMA4 optimizations: AMD processors |
38 |
starting with Bulldozer architecture. |
39 |
|
40 |
> sse4_1 - Enable SSE4.1 instruction support |
41 |
|
42 |
media-libs/libvpx:sse4_1 - Enable optimization for SSE4_1 capable |
43 |
processors (Intel Core 2 Penryn and later chips) |
44 |
|
45 |
|
46 |
> sse4_2 - Enable SSE4.2 instruction support |
47 |
|
48 |
media-video/ffmpeg:sse4_2 - Enables SSE4.2 optimizations: Nehalem-based |
49 |
Intel Core i7 or later. |
50 |
|
51 |
> ssse3 - Use the SSSE3 instruction set (NOT sse3/pni) |
52 |
|
53 |
media-libs/libvpx:ssse3 - faster floating point optimization for SSSE3 |
54 |
capable chips (Intel Core 2 and later chips) |
55 |
|
56 |
|
57 |
|
58 |
|
59 |
|
60 |
|
61 |
I like the "available in cpu $foo and later" because this influences me |
62 |
in what to buy: If I know that I will have noticeable gain with hand |
63 |
written asm in a package that is critical to me, I know it will perform |
64 |
better for my needs despite benchmarks might be saying otherwise. |
65 |
|
66 |
|
67 |
Alexis. |